Intelligent domain matching method and system based on intention large model

• Main Authors: WANG WEI, JIA HUIDI, ZHU SHANXIAN, CHANG PENGHUI, ZOU KEXU, KWAK DONG-SHIN, HUANG SI, SUN YUELI
• Format: Patent
• Language: Chinese; English
• Published: 15.03.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention relates to the field of natural language processing, in particular to a natural language semantic analysis method in a man-machine conversation system, in particular to an intelligent field matching method based on an intention large model, which trains the intention large model specially used for user intention recognition by constructing a prompt set, the prompt set comprises a wide range of user intention examples, and the intention large model is used for user intention recognition. Different fields, different industries and different application scenes are covered, and the model can better understand and recognize various user intentions. And meanwhile, by constructing a subdivided domain module and a task interface, a trained intention large model is utilized to match the problem input by the user with the corresponding description of the domain module and the task interface, so that a specific function which the user wants to realize can be accurately matched. 本申请发明涉及自然语言处理领域，尤其涉及人机对话系统中的